near 3D mis-calculation

[Anonymous User]

Original User: mwstbrk

I am trying to get the straight-line distance from points to lines along a change in z-distance (elevational change).  I believe the "Near 3D" tool accomplishes this.  However,  some of the algorithms obviously did not measure to the nearest lines (see attached).  All of the images are from the same map document, but please note the "correct" vs. "incorr" file names.

Thank you for any help in solving this problem!

Matt

[Anonymous User]

Original User: jbswain

If the lines are not of consistent elevation, then what you are seeing is possible.  The vertices on the line will be preferred for the snapping.  Also if you are using some type of arc rather than a poly line then that may also explain it, as the vertices would be selected not the curve.

[MatthewWestbrook]

Hello again and thank you!

In response to the following:

If the lines are not of consistent elevation, then what you are seeing is possible.  The vertices on the line will be preferred for the snapping.  Also if you are using some type of arc rather than a poly line then that may also explain it, as the vertices would be selected not the curve.

I have two shape files which I manually created by drawing new polylines.  Some of the polylines were drawn using vertices (clicking repeatedly to connect the dots) or using a trace technique (I held down the left mouse button and did not pick up the mouse----is this the "arc" that you were talking about?).  The lines are at different elevations, but not at great enough elevations to select the lines that were farther away.  Another problem might be that I created "polyline ZM" features, but my attribute table does not contain Z values (see attached). [ATTACH=CONFIG]27196[/ATTACH]

Do I need to create my polylines so they are compatible with the "Near 3D" tool? Or can I work with what I have?

[Anonymous User]

Original User: jbswain

The 3D lines will have the 'Z' values in the vertices as described in your collection process, they will not be in the attribute table.  For a test on the 'bad' locations, I would consider converting the lines to points at the vertices and see if the same vertices are selected.  Or if the lines do not vary much, then simply specify one elevation for them and then convert them to 3D via the Feature to 3D by attribute tool and then re run the test.  I don't think there is anything in the Near 3D algorithm that will only select unique vertices, but unless there is a specific change between the vertices in the line it sure looks like it. 

After the point test and the constant elevation test, I would consider contacting Esri Support for them to take a look at it directly for a potential bug report.

[MatthewWestbrook]

I think I accomplished the 3 tests.

Here is the original image after running "near 3d" from points to polylines:
[ATTACH=CONFIG]27247[/ATTACH]
I then converted the lines to points at their vertices and ran "near 3d" from points to the vertices:
[ATTACH=CONFIG]27246[/ATTACH]
I then changed the elevation of the vertices all to 0 by using "feature to 3d by attribute" and reran "near 3d":
[ATTACH=CONFIG]27248[/ATTACH]

The result using a uniform elevation=0 seems to provide the best result from a visual standpoint i.e., the results are what we expect.  However, this can't provide us with an accurate line measurement in 3D space which is what I am trying to measure.

Thanks again for the help.  Should I take this to ESRI support directly? I'm not sure what their contact information is.
-Matt

[Anonymous User]

Original User: jbswain

At elevation zero it appears to be working fine and selecting multiple connections to individual vertices, which I thought it was not doing. It appears to me that it is working, despite you not getting the results you are looking for.  The ones with the closest z elevation are being selected, since it seems you mentioned that the vertices are not the same.  You can contact Support via the Customer Care Portal or this page. I would consider the elevation of the points and then reconsider the results.  If you simply want the straight line then drop all the elevations to match, but with variable elevations, you will get different results.